Organize huge photo collections
Resize pictures, create previews and convert image batches without tying up your main desktop.
ANDROID COMPUTEMore power.
More power.
Less waste.
Turn compatible Android phones into a distributed processing fleet. Connect them to SamyStackOS over USB or Wi-Fi, and let a central coordinator put their processors to work on supported jobs—right alongside your mini PCs and Raspberry Pis.
In development · Explore the vision
A NEW ROLE FOR EXISTING HARDWARE
Big ideas don’t always need big machines.
Even a spare smartphone has a capable processor, usable memory and a network connection. Instead of leaving those resources idle, SamyStackOS is being designed to bring them together in one coordinated computing system.
Meet the pocket-sized compute fleet.
- Install the SamyStack Node app on supported Android phones
- Connect through a powered USB data hub or your local Wi-Fi
- Authorize each device with the Debian-based cluster controller
- Let the scheduler match jobs to available CPU, memory and supported acceleration
- Watch work progress across the fleet from one dashboard
9:40▴ ▰
~/samystack-worker
$ python worker.py
pixel-4a
✓ authenticated
arm64 | 4 cores
task #401
load model
34% complete
[worker] READY
9:41▴ ▰
~/samystack-worker
$ python worker.py
galaxy-s9
✓ authenticated
arm64 | 4 cores
task #402
tokenize batch
71% complete
[worker] READY
9:42▴ ▰
~/samystack-worker
$ python worker.py
pixel-3
✓ authenticated
arm64 | 4 cores
task #403
embed vectors
28% complete
[worker] READY
9:43▴ ▰
~/samystack-worker
$ python worker.py
oneplus-6
✓ authenticated
arm64 | 4 cores
task #404
resize tiles
86% complete
[worker] READY
9:44▴ ▰
~/samystack-worker
$ python worker.py
moto-g7
✓ authenticated
arm64 | 4 cores
task #405
hash objects
54% complete
[worker] READY
9:45▴ ▰
~/samystack-worker
$ python worker.py
galaxy-a50
✓ authenticated
arm64 | 4 cores
task #406
process audio
48% complete
[worker] READY
9:46▴ ▰
~/samystack-worker
$ python worker.py
pixel-5
✓ authenticated
arm64 | 4 cores
task #407
classify batch
92% complete
[worker] READY
9:47▴ ▰
~/samystack-worker
$ python worker.py
lg-v40
✓ authenticated
arm64 | 4 cores
task #408
run inference
66% complete
[worker] READY
9:48▴ ▰
~/samystack-worker
$ python worker.py
nokia-7
✓ authenticated
arm64 | 4 cores
task #409
queue task
79% complete
[worker] READY
9:49▴ ▰
~/samystack-worker
$ python worker.py
galaxy-s10
✓ authenticated
arm64 | 4 cores
task #410
upload results
40% complete
[worker] READY
ONE CONTROL CENTER
One dashboard. Every worker in view.
From a single screen, see which phones are connected, how much processing capacity they can offer and what each device is working on. Assign work intelligently and keep the whole fleet organized without juggling ten separate screens.
ANDROID WORKER FLEETINTERFACE PREVIEW
Galaxy S9CPU 48% · RAM 2.1 GBarm64 · Wi-FiReady
Pixel 4aCPU 26% · RAM 3.7 GBarm64 · Wi-FiReady
Moto G PowerCPU 54% · RAM 1.8 GBarm64 · Wi-FiWorking
Galaxy A32CPU 12% · RAM 3.2 GBarm64 · Wi-FiReady
LG V40CPU 69% · RAM 2.4 GBarm64 · Wi-FiWorking
Pixel 3aCPU 0% · Charging requiredarm64 · standbyPaused
LIVE FLEET MANAGEMENT
Stay in control of every device.
The planned dashboard brings connected phones, workload progress and hardware health into the same place as the rest of your SamyStackOS cluster.
- Connected, available, working and paused workers
- CPU load, available memory and temperature monitoring
- Per-phone processing and charging preferences
- Supported workload types and scheduling limits
- Completed jobs, errors and retry history
WHAT COULD YOUR FLEET DO?
Small devices. Surprisingly useful jobs.
Imagine splitting a big job into many smaller tasks, then letting available phones tackle those tasks at the same time. Each device works within its own capabilities while SamyStackOS coordinates the results.
Work through file backlogs
Check file integrity, inspect archives and divide compatible background processing between multiple phones.
Run lightweight AI tasks
Spread small-model inference, classification and data preparation across phones that support the required runtime.
Process media in parallel
Handle compatible audio clips or video segments as independent jobs, with hardware-aware limits.
Speed through repetitive data work
Divide calculations and data batches into smaller tasks that many processors can complete together.
Make it your own
Extend the system with approved task modules built for the Android worker runtime.
HOW ANDROID JOINING WORKS
Plug in. Pair up. Put it to work.
The Debian coordinator manages the job queue while each Android phone runs its own lightweight worker. No complicated manual assignment is needed for every task.
Connect
Install the Android app and connect the phone by USB data cable or local Wi-Fi.
Authorize
Approve the new worker in SamyStackOS and set its resource and charging limits.
Compute
The scheduler sends supported tasks to eligible phones and tracks progress.
Complete
Results return to the coordinator, ready for the next stage of the workload.
FLEXIBLE CONNECTIONS
Built for the shelf, the desk or a full phone fleet.
Phones can stay in one place and work together, or join over your home network. SamyStackOS is designed to find the best available connection without making you manage every device separately.
One cable. Two jobs.
A compatible powered USB hub can keep phones charged while providing a wired data connection to the Debian host. A neat, predictable setup for permanently mounted worker phones.
Same cluster. No fixed location.
Phones can also communicate over local Wi-Fi, making it easy to add another available device without rebuilding your hardware setup.
DESIGNED FOR REAL DEVICES
Useful power, without burning out your phones.
Each worker is designed around practical limits: temperature, battery health, charging preferences, memory availability and the ability to pause or disconnect at any time.
Development status: Android cluster computing is a planned SamyStackOS capability, and the Android APK is not yet available for download. Each phone keeps its own memory; the coordinator distributes compatible jobs rather than combining every device into one physical computer. Performance and supported tasks will vary by device.